Neidhardt, MaximilianMaximilianNeidhardtBhattacharya, DebayanDebayanBhattacharya2026-07-022026-07-022026-07-02https://hdl.handle.net/11420/63594This dataset contains synchronized stereo-camera RGB recordings of zebrafish swimming with varying group sizes (2, 5, 10 fish), along with manually annotated frames for tracking and evaluation. It supports training and benchmarking of deep learning–based multi-object tracking and classification. The dataset includes raw video data, annotations, model weights, and example predictions. The data processing is described in detail in: Neidhardt, Maximilian, et al. "Robust Motion Tracking and Classification of Zebrafish with Deep Learning." IEEE Access (2026).enhttps://creativecommons.org/licenses/by-nc/4.0/Deep learningFish trackingFish trajectory datasetLocomotion patternPhysical activityTechnology::610: Medicine, HealthComputer Science, Information and General Works::006: Special computer methods::006.3: Artificial IntelligenceVideohttps://doi.org/10.15480/882.1735010.15480/882.17350Neidhardt, MaximilianMaximilianNeidhardtLatus, SarahSarahLatusUniversitätsklinikum Hamburg-Eppendorf (UKE)